Date: Fri, 18 Apr 2003 12:40:13 -0700 (PDT) From: Gary Hanson <garyh@agora.rdrop.com> To: freebsd-bugs@FreeBSD.org Subject: Re: i386/40274: "fxp: device timeout" errors during heavy,sustained network activity Message-ID: <200304181940.h3IJeDne018524@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR i386/40274; it has been noted by GNATS. From: Gary Hanson <garyh@agora.rdrop.com> To: freebsd-gnats-submit@FreeBSD.org, seagull@aracnet.com Cc: Subject: Re: i386/40274: "fxp: device timeout" errors during heavy, sustained network activity Date: Fri, 18 Apr 2003 12:33:53 -0700 (PDT) Also seen on 4.8-RELEASE and 5.0-RELEASE; very easy and quick to reproduce. System is an Intel N440BX dual proc mobo with integrated 82557 (Pro100B) NIC on a dedicated IRQ. On SMP kernels, I could ping for as long as I wanted to, even with large (64000 byte) packets. Didn't try flood ping though. But try to do some real I/O like an ftp or Samba file copy and the transfer would die almost immediately and the console would start reporting fxp: device timeout messages every few seconds, even after disconnecting the ethernet cable. The system was then in a very weird state; it was acting like it was very busy (sluggish) but wasn't totally unresponsive, and it could not be shutdown properly - that would seem to hang (or I wasn't willing to wait long enough). With a single-processor kernel, 4.8 and 5.0 both had no problems. The hardware is good; it works fine with SMP OSes from Redmond; I'll try a penguin 'OS' for amusement purposes. [The penguin 'e100' driver is vastly better than the rather old FreeBSD fxp driver, but I'm not competent to adapt/port/steal it.]
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200304181940.h3IJeDne018524>